Multimodular polyketide synthases (PKSs) are enzymatic assembly lines that synthesize many antibiotics. We seek to understand and engineer such systems, using the 6-deoxyerythronolide B synthase (DEBS) as a prototype. This revised application addresses the reviewers' major concerns by focusing primarily on two aspects of DEBS that are arguably of greatest interest to biosynthetic engineers: programmed unidirectional channeling of intermediates along the assembly line, and extender unit specificity. Our Specific Aims are: Aim 1: Reconstitution and analysis of DEBS: Although DEBS has been studied for two decades, in vitro reconstitution and analysis of the entire system has not been reported. Kinetic analysis of 6-deoxyerythronolide B synthesis will not only lay the quantitative foundation for Aims 2 and 3, but will also answer basic questions of broad relevance to all of nature's PKS assembly lines. Aim 2: Understanding and engineering protein-protein interactions in assembly line PKSs: Our studies have suggested that three types of non-covalent protein-protein interactions can collectively explain how PKS assembly lines have been evolutionarily programmed for unidirectional chain growth. We will perform two sets of experiments to test this hypothesis. First, we will construct and analyze mutants in which the directionality of polyketide chain growth is altered. Second, by directed manipulation of protein-protein interactions, we will improve the efficiency of representative chimeric PKSs that show measurable but sub-optimal turnover. Aim 3: Understanding and engineering extender unit specificity of assembly line PKSs: Acyl transferase (AT) domains catalyze extender unit incorporation in each module of an assembly line PKS. Although a few cases of engineering AT specificity via homology-based approaches have been reported, the resulting PKSs are kinetically impaired and the underlying causes are unclear. We propose a fundamentally different strategy to engineer AT specificity. First, we will investigate the mechanistic basis for acyl-CoA and ACP specificity via kinetic analysis of three AT domains whose structures we have solved. The resulting methods and insights will enable quantitative interrogation of AT mutants whose design is guided by either structure-based or statistics-based principles. Finally, by re-introducing the most promising AT mutants into the DEBS assembly line, we will evaluate the consequences of AT domain engineering on the turnover rate of the full system. Aim 4: Preparation and evaluation of 15-propargyl-erythromycin A analogues: While not directly related to our primary future directions, during the past funding cycle we discovered 15-propargyl erythromycin A, a unique orthogonally functionalized antibiotic with similar activity to erythromycin A. With modest effort, we propose t prepare a few derivatives for ongoing studies on macrolide-ribosome interactions in our collaborator's lab. Our studies will not only open new doors for biosynthetic engineers who seek to manipulate antibiotic structure by genetic engineering, but will also highlight the broader relevance of polyketide biosynthetic research.